Skip to content
Brand Stack
Esc
↑↓navigate↵open⌘Jpreview
Sign up
On this page

Version History, Proposals & Connections

Every edit to a Brand Stack cuts a version. See what changed and when, review incoming change proposals from connected repos, and track which codebases are running your brand.

A Brand Stack isn’t a one-time export anymore — it’s a living record that tracks its own history and stays in sync with the codebases that use it.

Every change cuts a version

Editing a section, regenerating an asset, or approving a proposal all produce a new, immutable, semver-numbered version of the stack (1.2.0 becoming 1.3.0, for example). Nothing is silently overwritten — you can always see what the stack looked like at any point and what changed between versions.

Version History

The History card in the Build panel (the right-hand column of the stack detail page) lists every version, newest first, with its release type and a changelog you can expand. The version badge in the page header links straight down to it.

Proposals: changes from your codebase, reviewed before they land

When an AI coding agent notices your brand has drifted from what’s in a connected repo — a retuned color, updated copy — it can file a change proposal instead of editing the stack directly. Proposals show up in the Proposals card with a field-by-field diff (old value → new value) so you can see exactly what’s being asked for.

  • Approve applies the change and cuts a new version. If part of a proposal conflicts with an edit you made directly in Brand Stack, that field is protected and skipped — the response tells you which paths landed and which were kept.
  • Decline closes the proposal without touching the stack.

A new proposal from the same repo automatically supersedes that repo’s earlier open proposals, so you’re never reviewing stale requests.

Connections: where your brand is deployed

The Connections card is a map of every repo your Brand Stack is connected to — populated automatically when an agent adopts a repo, files a proposal, or runs brandstack pull. Each row shows the repo, its detected target profile, which version it’s pinned to, and when it was last seen, so you always know whether a codebase is running your current brand or has drifted behind.

Bringing a repo into Brand Stack

An AI coding agent working inside a codebase can adopt that repo as a new Brand Stack, or connect it to an existing one, without leaving the terminal:

  • New stack — the agent analyzes your styles, fonts, copy, and README, and Brand Stack generates a complete stack from what it observed.
  • Existing stack — the agent submits its observations as a change proposal for you to review in the Proposals card above, rather than applying anything automatically.

Was this page helpful?